Validated Results

Results

Hardware measurements from a Wi-Fi 7 testbed and Monte Carlo simulations confirming that CCI-based adaptive switching outperforms fixed ZF and fixed MRT strategies.

Key Findings

↓41%

Outage rate

WaveLynk vs. fixed ZF

↓32%

Latency

at weak signal (−80 dBm)

↓30%

Packet loss

WaveLynk vs. fixed ZF

100

Trial count

hardware + Monte Carlo

Comparison Summary

WaveLynk vs. Fixed Strategies

Measured across 100 independent trials, spanning weak to strong signal conditions.

Metric	Always ZF	Always MRT	WaveLynk (Adaptive)	Improvement vs. ZF
Outage rate (%)	~28%	~12%	~17%	↓41%
P50 latency (ms) at −80 dBm	~65	~50	~44	↓32%
Packet loss at −80 dBm (%)	~45%	~18%	~5%	↓89%
Peak throughput (−50 dBm)	High	Moderate	High (ZF preserved)	—
Failure mode	Sudden collapse	Gradual degradation	Prevented	—

Values from 100 hardware trials on Wi-Fi 7 testbed. See Research for full methodology.

Hardware Validation

Experimental results — Wi-Fi 7 testbed

Measured using iperf3 on a Netgear Nighthawk Wi-Fi 7 router at 6 GHz, with 2 client laptops and 2 monitor smartphones across 100 independent trials.

Latency over time: Fixed ZF vs WaveLynk — latency stabilizes after switch

Fig. 1 — Latency over time. Fixed ZF latency rises unbounded as channel conditions degrade. WaveLynk switches to MRT at t ≈ 22 s, immediately stabilizing latency to ~44 ms. Error bars: n = 5 repetitions.

Latency and packet loss vs received signal power — WaveLynk vs Fixed ZF

Fig. 2 — Latency and packet loss vs. received signal power. At weak signal (−80 dBm), fixed ZF collapses while WaveLynk maintains stable performance. Both strategies converge at strong signal where CCI stays below γ.

Ping latency vs received signal power — MRT-dominant vs ZF-dominant

Fig. 3 — Ping latency vs. signal power. MRT-dominant mode (WaveLynk when CCI ≥ γ) consistently outperforms ZF-dominant mode at signal levels below −65 dBm — the coherence cliff transition zone.

Packet loss rate vs received signal power — MRT-dominant vs ZF-dominant

Fig. 4 — Packet loss rate vs. signal power. At −80 dBm, ZF-dominant mode shows 45% packet loss vs. 18% for MRT-dominant. WaveLynk's adaptive strategy captures the MRT benefit exactly when needed.

Methodology

Hardware testbed setup

Equipment

Router: Netgear Nighthawk Wi-Fi 6/7, 6 GHz band
Server: 1 laptop running iperf3 as throughput server
Clients: 2 laptops running iperf3, stressing MU-MIMO
Monitors: 2 smartphones measuring live RSSI and throughput
Trials: 100 independent measurement runs

Experimental controls

Same client device and router for all trials
Fixed distance and antenna orientation per trial
Same channel and bandwidth throughout
Identical traffic pattern (iperf3 UDP stream)
Each measurement repeated n = 5 times; error bars shown
Signal strength varied by physical positioning

Comparison protocol: each signal strength condition was tested under fixed ZF mode (ZF at all times), fixed MRT mode (MRT at all times), and WaveLynk adaptive mode (CCI-based switching with γ = 0.6). Strategies used identical hardware and channel conditions — only the precoding mode differed.

Overview

Project poster

Summary poster from the WaveLynk science fair presentation, showing the core discovery, experimental results, and adaptive switching architecture.

WaveLynk — Charting a New Frontier in 6G Reliability. Overview poster showing the core insight: adaptive beamforming expands the stable 6G operating region by 30%, preventing the abrupt link collapse characteristic of fixed ZF strategies.

Dig deeper

Read the paper for full derivations, or run the simulation notebooks yourself.

View the Research → Try Live Models